Design and Modeling of a Parallel-Pipe-Crawling Pneumatic Soft Robot

Zhiyuan Zhang, Xueqian Wang, Songtao Wang, Deshan Meng, Bin LiangIEEE Access 2019 · DOI

Introduction

Inspecting industrial piping networks is a critical yet hazardous task. Traditional rigid robots, while robust, often struggle with the geometric constraints of tight bends and risk damaging fragile pipe interiors. Soft robotics offers a compelling alternative due to inherent compliance; however, existing "series-structured" soft robots—which mimic the peristaltic motion of earthworms—are typically limited by low crawling speeds and a lack of active steering capabilities.

In a recent study published in IEEE Access, Zhang et al. present a novel pneumatic soft robot that addresses these limitations. By employing a parallel-actuator design, the researchers achieved record-breaking crawling speeds and the ability to navigate complex three-dimensional pipe environments.

Robot Design and Parallel Architecture

The robot's architecture departs from the traditional linear series model in favor of a parallel structure. It consists of three extensible pneumatic soft actuators (ePSAs) arranged in a triangular configuration, secured by intermediate connecting rings. This arrangement provides the robot with three degrees of freedom: pitch, yaw, and extension, enabling active steering at T-junctions and bends.

The parallel structure of the soft robot, featuring three ePSAs, connecting rings, and directional feet.

Locomotion is facilitated by "flexible feet" attached to the robot's ends. These feet utilize directional friction; a smooth frontal surface and a grooved rear surface create a passive ratcheting effect. As the ePSAs inflate and deflate, the robot effectively "walks" forward through the pipe without requiring complex active anchoring mechanisms.

Actuator Modeling and Optimization

To ensure precise control without the computational overhead of complex non-linear elasticity equations, the team developed a phenomenological "three-element model." This model treats the soft actuators as a combination of thrust, spring, and damping elements. Experimental validation showed this simplified approach to be highly accurate, with an average length error of only 0.51 mm.

Structural diagram of a single extensible pneumatic soft actuator (ePSA) showing the fiber-wound elastic body.

The design of the individual ePSAs was optimized using Finite Element Method (FEM) simulations in Abaqus. The researchers identified optimal parameters—such as an air chamber radius of 3 mm and 20 fiber winding turns—to maximize longitudinal extension while minimizing radial expansion (ballooning).

Experimental Results

The parallel-structured robot demonstrated significant performance improvements over existing soft pipe-crawlers:

  • Speed: It achieved a maximum horizontal crawling speed of 18.3 mm/s, the highest recorded for peristaltic soft robots in the literature reviewed.
  • Load Capacity: The robot can support a theoretical maximum load of 2.456 kg, allowing it to carry necessary inspection equipment like endoscopic cameras.
  • Maneuverability: The robot successfully navigated 90-degree L-shaped bends and flexible bellows with a minimum turning radius of 38.2 mm. It also demonstrated vertical climbing capabilities at speeds of 12.9 mm/s.

Conclusion

The parallel-actuator design represents a significant leap in soft robot efficiency and maneuverability for industrial pipe inspection. By combining a simplified phenomenological model with optimized hardware, the researchers created a fast, strong, and agile system. Future work will focus on integrating internal flexible sensors to transition from open-loop to closed-loop control, further enhancing the robot's adaptability in unpredictable environments.

Summary updated

BibTeX
@article{zhang2019design,
  author = {Zhiyuan Zhang and Xueqian Wang and Songtao Wang and Deshan Meng and Bin Liang},
  title = {{Design and Modeling of a Parallel-Pipe-Crawling Pneumatic Soft Robot}},
  journal = {IEEE Access},
  year = {2019},
  doi = {10.1109/access.2019.2941502},
}
Figure